Detection of Adversarial Attacks in Robotic Perception
This work tackles safety risks in robotics by improving adversarial attack detection, but it appears incremental as it builds on existing robustness research for image classification.
The paper addresses the vulnerability of Deep Neural Networks in semantic segmentation for robotic perception to adversarial attacks, proposing specialized detection strategies to enhance safety in critical applications.
Deep Neural Networks (DNNs) achieve strong performance in semantic segmentation for robotic perception but remain vulnerable to adversarial attacks, threatening safety-critical applications. While robustness has been studied for image classification, semantic segmentation in robotic contexts requires specialized architectures and detection strategies.